[AccessD] Handling concurrent updates

Bill Benson bensonforums at gmail.com
Sat Feb 6 07:53:32 CST 2016


I am struggling to set up a test of performance using Gustav's process,
which I believe keeps his cursor on each non-editable record until it
becomes editable, versus mine, which puts non-editable records into a
collection and comes back to them after all editable records have been
edited. First off, it is hard to think of a test that will cause the kind
of conflict I want to monitor "in the laboratory."


I thought of letting 1 process have a form bound to a table, and a second
instance with a DAO recordset that is opened containing the same records.
Suppose the user is on record 1, and the 2nd process is looping through a
recordset beginning at record 1.

Using Gustav's method, the recordset using process in instance 2 will find
record 1 uneditable. If the edits to be made to record 2 are made to wait
until the form in process 1 releases record 1, then record 2 will not be
updated at the time the user operating via the form leaves record 1 and
goes to record 2.


Using my proposed method (as yet untested), the process using the recordset
would have put record 1 in a collection for later treatment; wuold have
gone on to record 2, updated data there... and when the user using the Form
in instance 1 navigates to record 2, lo and behold, all the data is as up
to date as it was required to be according to logic that drove instance 2
to make those allowed edits.

Personally I think this is safer.


OMMV.


And as I have said over and over - I have not tested this to see if it
could even work (because I cannot think how to set it up LOL).


More information about the AccessD mailing list